Open tonyandsunny opened 4 years ago
Except for the pre-processing part (which is pure numpy), this model can be easily implemented with PyTorch Geometric (e.g. see mnist_graclus.py
). The pre-processing segmentation/adjacency/feature computation can be found here. After that, you just need to convert it to the PyG format :)
Thank you very much for your work, I am not very familiar with the tensorflow version, do you have this code for pytorch?